Growth: some combined cross- sectional and time series evidence from OECD countries

The uncovered interest rate parity hypothesis and three variants of the monetary approach to exchange rate determination are assessed under a vector autoregression representation of the available information variables, using monthly data on six major US dollar exchange rates over the period 1978–90. A large information set is used, and the time series properties of the information variables are taken into account. The cross-equation restrictions imposed on the estimated parameters are tested statistically and the economic significance of the models is evaluated independently on the basis of appropriate volatility tests. A weak test for exchange rate bubbles, based on a decomposition of market noise, is proposed.